U spec templates
Skill zig999/siegard-code/dist/.claude/skills/u-spec-templates
Most AI coding tools help you write code. Siegard Code manages the entire development lifecycle — it writes specifications, plans backlogs, implements features, runs QA, and delivers tested code. All autonomously, all traceable, all through Claude Code.
npx -y skills add zig999/siegard-code --skill u-spec-templatesAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 9 stars9 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Canonical TEMPLATE.* artifacts for SDD spec writers — domain spec, back spec, front spec, feature spec, flow, component spec, decisions, design-system rules and design-system bundle. Consumed by u-spec-writer, u-spec-back, and u-spec-front, which read templates by path. Also ships scripts/read_spec_sections.py, which loads only the sections a worker needs while always returning the full section index (R16). Not user-invocable.
SKILL.md
3.8 KB, 876 tokens by cl100k_base, as published. Nobody here has run it
u-spec-templates
Canonical templates consumed by SDD spec agents, read by path (.claude/skills/u-spec-templates/<file>); the directory listing is authoritative.
scripts/read_spec_sections.py
Loads the sections a worker needs and returns the complete section index either way, so partial loading never becomes partial awareness.
python3 .claude/skills/u-spec-templates/scripts/read_spec_sections.py \
--file "$SPECS_DIR/domains/<domain>/back/<domain>.back.md" \
--sections "Business Rules,State Machine" # numbers, §N, or title text
# --index-only : titles and line counts, no bodies
# --all : explicit opt-out of scoping
# exit 2 : a selector matched nothing (reported, never silently dropped)
Sections are level-2 headings; the shipped templates number them (## 4. State Machine (ST)), and
unnumbered ones (## Changelog) are addressable by title. The preamble before the first heading is
always included — it carries the artifact's identity and version.
Measured effect (troubleshooting-engine, the domain that consumed 57,213 of a 60,000 token
budget):
| Worker | Artifact | Whole file | Scoped |
|---|---|---|---|
u-spec-back | .spec.md | 1046 | 833 (−20%) |
u-spec-front | .spec.md | 1046 | 259 (−75%) |
u-spec-validator | .spec.md | 1046 | 833 (−20%) |
u-spec-validator | .back.md | 1374 | 1168 (−15%) |
u-spec-reviewer | .spec.md | 1046 | 1046 (0% — by design) |
The back/validator saving is modest because §Business Rules is the largest section (763 of 1374
lines in that .back.md) and those workers genuinely need it. u-spec-front is the large win: it
needs UI-facing states and errors, not backend enforcement detail.
u-spec-reviewer and u-spec-compliance read whole files deliberately — one checks completeness,
the other scans for gaps, and both are defeated by a subset.
Index
| Template | Produces | Primary consumer |
|---|---|---|
TEMPLATE.spec.md | domains/{domain}/{domain}.spec.md | u-spec-writer |
TEMPLATE.back.md | domains/{domain}/back/{domain}.back.md | u-spec-back |
TEMPLATE.front.md | {SPECS_DIR}/front/front.md (global frontend spec) | u-spec-front |
TEMPLATE.feature.spec.md | {SPECS_DIR}/front/features/*.feature.spec.md | u-spec-front |
TEMPLATE.flow.md | {SPECS_DIR}/front/flow.md | u-spec-front |
TEMPLATE.component.spec.md | design-system/components/*.spec.md | u-fe-spec-writer |
TEMPLATE.decisions.md | {SPECS_DIR}/decisions.md | u-spec-writer |
TEMPLATE.design-system-rules.md | front/design-system-rules.md | u-spec-front |
TEMPLATE.design-system/ | design-system/ bundle (_index.md, tokens.md, composition.md, components.md, implementation.md) | u-spec-front |
FRONTEND-MANDATORY-ARTIFACTS.md | single source of truth for the frontend design-system artifacts the front pipeline must produce and the validator blocks on (F-07) | u-spec-front (produces), u-spec-validator (gates) |
Constraints
- Templates contain
<!-- INSTRUCTION: ... -->placeholders — producers MUST resolve every placeholder; none may survive into generated artifacts - Identifier prefixes used across templates follow the global pattern defined in
u-spec-globals/conventions.md(UC, BR, ST, EV, UI, FL)
What ships with it: 15 files
106.9 KB alongside SKILL.md, 1 of them executable
TEMPLATE.design-system/
- components.md12.4 KB
- composition.md10.5 KB
- implementation.md13.5 KB
- _index.md2.7 KB
- tokens.md17.1 KB
scripts/
- read_spec_sections.pyruns7.3 KB
- FRONTEND-MANDATORY-ARTIFACTS.md1.9 KB
- TEMPLATE.back.md4.1 KB
- TEMPLATE.component.spec.md6.1 KB
- TEMPLATE.decisions.md630 B
- TEMPLATE.design-system-rules.md9.1 KB
- TEMPLATE.feature.spec.md7.9 KB
- TEMPLATE.flow.md4.6 KB
- TEMPLATE.front.md5.9 KB
- TEMPLATE.spec.md3.1 KB
Gives 0 of the 12 instructions most plan spec skills give in 876 tokens
Counted across 1,099 of the 1,860 authors here whose files we hold, read 2026-08-07
- Ask one question at a timein 51 of 1099
- Break plans into vertical slicesin 29 of 1099, across 11 files
- Publish issues in dependency orderin 27 of 1099, across 9 files
- Iterate until user approves the breakdownin 25 of 1099, across 7 files
- Explore the repository to understand the codebase statein 24 of 1099, across 7 files
- Use domain glossary vocabularyin 23 of 1099, across 5 files
- Apply correct triage labels to published issuesin 23 of 1099, across 5 files
- Prefer AFK slices over HITLin 22 of 1099, across 7 files
- Write a specification before writing any codein 22 of 1099, across 14 files
- Write failing tests before implementation codein 22 of 1099, across 20 files
- Ask clarifying questions until requirements are concretein 21 of 1099, across 13 files
- Respect existing architecture decision recordsin 20 of 1099, across 5 files
Said here and by no other author read
- read templates by path
- return the complete section index
- include the preamble before the first heading
- drop repository paths and identifiers
Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.